Error in model prediction: zero-size array to reduction operation maximum which has no identity. Please check the model.

[wwmpython]

[CVHub520]

Hi there! @wwmpython, 😃

Thank you for reporting the issue. To help me diagnose and replicate the problem, could you please package the ONNX file, the configuration file, and a sample image that you're certain the model should be able to predict the target on into a compressed archive?

Please send the compressed package to the following email address for further investigation: cv_hub@163.com. Your assistance in providing these files is crucial for us to resolve the issue. Thank you in advance for your cooperation!

[wwmpython]

Hi there! @wwmpython, 😃

Thank you for reporting the issue. To help me diagnose and replicate the problem, could you please package the ONNX file, the configuration file, and a sample image that you're certain the model should be able to predict the target on into a compressed archive?

Please send the compressed package to the following email address for further investigation: cv_hub@163.com. Your assistance in providing these files is crucial for us to resolve the issue. Thank you in advance for your cooperation!

ok

[CVHub520]

Hey there, @wwmpython! After reviewing the files you provided, I noticed an issue with your ONNX file:

It appears there's a mistake that needs correcting. Please re-export your *.pt file using the official YOLOv8 repository. You can follow the guidelines in this tutorial for the export process.

If you continue to experience issues, don't hesitate to reach out for further assistance.

[wwmpython]

当我按照官方的代码将我的权重文件转换为onnx模型时，出现如下错误：Traceback (most recent call last):
File "F:\wwm\yolov8-pytorch-master\onnx_trans.py", line 10, in
model = YOLO("F:\wwm\yolov8-pytorch-master\model_data\yolov8_s.pt")
File "F:\anaconda3\envs\onnx\lib\site-packages\ultralytics\models\yolo\model.py", line 23, in init
super().init(model=model, task=task, verbose=verbose)
File "F:\anaconda3\envs\onnx\lib\site-packages\ultralytics\engine\model.py", line 149, in init
self._load(model, task=task)
File "F:\anaconda3\envs\onnx\lib\site-packages\ultralytics\engine\model.py", line 230, in _load
self.model, self.ckpt = attempt_load_one_weight(weights)
File "F:\anaconda3\envs\onnx\lib\site-packages\ultralytics\nn\tasks.py", line 857, in attempt_load_one_weight
model = (ckpt.get("ema") or ckpt["model"]).to(device).float() # FP32 model
KeyError: 'model'

进程已结束,退出代码1

[CVHub520]

I highly recommend utilizing the official repository (https://github.com/ultralytics/ultralytics) for training and exporting purposes.

Upon reviewing the image you provided, it is evident that you are using a third-party repository, which is not advisable.

If you still have compelling reasons that necessitate using a non-official repository, I urge you to write your own export code to ensure consistency with the official standards.